Social Networks and Health

Social Networks and Health Lisa F. Berkman, Ph.D. Cabot Professor of Public Policy and Director, Harvard Center for Population and Development Studies WHO Geneva, June 2-4,2010

Social Network Model IND. Personal Cell Intimate Zone A Intimate Zone B Effective Zone Nominal Zone Extended Zone Source: Bolssevain, Jeremy: Friends of Friends, 1974

Conceptual Model of How Social Networks Impact Health Upstream Factors Downstream Factors Condition: the extent, shape, and nature of Which provides opportunities for Which impacts health through the following Social-Structural Conditions (Macro) Social Networks (Mezzo) Behavioral Mechanisms (Micro) Pathways Culture Language Norms and values Competition/cooperation Socioeconomic Factors Inequality Discrimination Conflict Labor Market structures Sociogeographic Factors Urban/rural Neighborhood characteristics Residential and occupational segregation Social Change Urbanization War/civil unrest Economic depression Network Structure Size Density Reciprocity Reachability Proximity Organizational structure Social ranking Network Activation Frequency of face-to-face interaction Frequency of nonvisual contact Frequency of organizational participation (attendance) Duration and intensity o Social Support/conflict Instrumental and financial Informational. Appraisal Emotional Conflict/negative Access to Resources and Material Goods Jobs/economic opportunity Access to health care Housing Human capital Referrals/institutional contacts Social Engagement Physical/cognitive exercise Reinforcement of meaningful social roles Interpersonal attachment Social Influence Constraining/enabling influences on health behaviors Attitudes and norms toward help-seeking Attitudes and norms toward treatment adherence Psychobiological Pathways Stress-response/allostatic lode Immune system function DHEA levels Glucocoticoid levels Blood pressure Cardiovascular reactivity inflammattion Health Behavioral Pathways Smoking/alcohol consumption Diet Exercise Adherence to medical treatments Help-seeking behavior Psychosocial Pathways Self-efficacy Coping effectiveness Relaxation/stress management Depression/distress Sense of well-being/qol

The upside of social integration, social support: experimentally induced support or disease exposure

Pittsburgh Common Cold Study N=276: Cohen et al JAMA,1997 VIRUS 6 Day Quarantine Social Roles Clinical Colds

Social Roles and Colds Cohen et al JAMA,1997 6 0 O R = 4.2 3 * % with Colds 5 0 4 0 O R = 1.8 7 * O R = 1.0 0 3 0 1-3 4-5 6 + # o f H ig h - C o n t a c t R o le s ( O d d s r a tio s a r e a d ju s te d fo r c o n tr o l v a r ia b le s.)

Social Support and Blood Pressure Reactivity to Challenge: experimentally induced blood pressure reactivity

Social Networks and Health: results from observational studies

Social integration and mortality in a French occupational cohort: EDF-GDF employees 3 2.5 2 1.5 1 high med/high med/low low 0.5 0 men women Adjusted for age, occupational grade, cigarette smoking, alcohol consumption, BMI, depressive symptoms, self-rated health, and geographical region.

Survival curves by marital status and 1 0.9 presence of close confidant % Survival 0.8 0.7 0.6 0.5 0 1 2 3 4 5 6 Year of Follow-up Married and/or confidant Unmarried, no confidant Williams et al., JAMA 1992

6-Month Mortality post MI by Emotional Support: EPESE 100 80 Relative Risk (%) 60 40 20 0 *P.05. 53% 36% 23% n = 53 n = 111 n = 26 0 Sources 1 Source 2 or More Sources Emotional Support* Source: Berkman, Leo-Summer, Horwitz: Emotional Support and Survival Following Myocardial Infarction. Ann Intern Med, 1992

Caregiving: the upside and downside

Caregiving in the Nurses Health Study. Lee,Colditz,Berkman,Kawachi, AM J Prev Med 2003:24(2):113-119 54,412 women in the Nurses Health Study, ages 46-71 (no documented CHD) Information on caregiving in 1992 CHD follow up 1992-1996 321 incident cases, 231 nonfatal, 90 fatal

CHD Risk: Caregivers of Disabled/Ill Spouse 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 297 8 16 Total CHD Hours of Caregiving Per Week: 214 7 10 0 1-8 >=9 Nonfatal CHD

Mounting evidence that caregiving takes a toll Schulz and Beach. Caregiving strain increases mortality risk ( RR=1.63) Caregivers have poorer immune function,slower at would healing( Kiecolt-Glaser et al) Caring for disabled child, telemere length and telemerase ( Epel, Blackburn et al) Caring for cognitively ill is especially stressful(epese)

Social integration predicts the preservation of memory in US elderly: Findings from the Health and Retirement Study Karen A. Ertel M. Maria Glymour Lisa F. Berkman

Methods: Study Sample Health and Retirement Study (HRS) Nationally-representative sample of US residents born before 1948 4 waves of data collection: 1998, 2000, 2002, 2004 N=16,638

Methods: Exposure Measure Social integration at baseline (1998) 1. Marital status 2. Volunteer activities 3. Contact with parents 4. Contact with children 5. Contact with neighbors Possible scores: 0, 1, 2, 3, 4, 5

Results: Trajectory of memory scores comparing subjects with high versus low social integration 10.6 Memory score 10.4 10.2 10.0 9.8 9.6 9.4 9.2 9.0 High social integration Low social integration 1998 2000 2002 2004 Year Predictions are for white males with mean age, income, wealth, and education, and healthiest baseline scores for all health measures.

MIDUS Brief Test of Adult Cognition by Telephone (BTACT) (Lachman & Tun, 2008) TASK THEORETICAL CONSTRUCT(S) SAMPLE ITEMS Word List Recall (immediate and delayed) Backward Digit Span Category Verbal Fluency Episodic verbal memory 15 words (e.g., flower, truck, school) Working memory Spans from 2 to 8 (e.g., 7, 2, 5, 3 = 3, 5, 2, 7) Executive function, semantic memory List as many animals as possible in 60 seconds Number Series Inductive reasoning 5 series of 5 numbers (e.g., 35, 30, 25, 20, 15 correct answer = 10) Backwards Counting Attention- Switching, Stop and Go Task Switch Processing speed Reaction time, attention, task switching Counting backward quickly from 100 Normal condition: Green = Go; Red = Stop Reverse condition: Green = Stop; Red = Go With mixed and switched trials

Executive Function & Social Conflict [avg. M1&M2] Age stratified Means 0.5 0-0.5 * * *** -1-1.5 32-44 (p=0.08) 45-54 (p=.002) 55-64 p=0.03) 65-74 (p=0.32) 75+ (p=0.13) Q1 Q2 Q3 Q4

Can we intervene to promote health by improving social networks and social support?

ENRICHD-Enhancing Recovery in coronary heart disease: Objective Berkman et al: Effects of Treating Depression and Low Perceived Social Support on Clinical Events After Myocardial Infarction: ENRICHD Randomized Trial. JAMA, 2003. 289(23) To test the hypothesis that treatment of depression and low social support early after an acute myocardial infarction will reduce death and nonfatal recurrent infarctions.

Study Design ENRICHD JAMA 2003;289:3106 Randomized, parallel-group clinical trial to compare the efficacy of a psychosocial intervention vs. usual care 2,481 post-mi patients with depression or low social support Average 3.4 years of follow-up Blinded ascertainment of primary endpoint Intention to treat analysis

Baseline to 6-month Changes in Social Support and Depression 10 5 5.1 Intervention effect p<0.001 0 Intervention -5-10 -15 ESSI -10.1 Hamilton depression score ESSI reported for patients with low social support only Hamilton depression score reported for depressed patients only

Baseline to 6-month Changes in Social Support and Depression 10 5 5.1 3.4 Intervention effect p<0.001 0-5 Intervention Usual care -10-15 ESSI score -10.1-8.4 Hamilton depression score ESSI reported for patients with low social support only Hamilton depression score reported for depressed patients only

Kaplan-Meier Survival Curves